Multi-language repository task suite (sample listing)
Illustrative listing: held-out repository-level tasks with executable acceptance tests for evaluating coding agents.
Sample listing. This page illustrates how a reviewed dataset is presented. It is not real inventory and cannot be licensed.
This sample listing illustrates an evaluation suite intended to stay out of training corpora.
Dataset facts
- Version
- v0.1-demo
- Records
- Not yet verified
- Coverage
- Not yet verified
- Languages
- Python, TypeScript, Rust, Go
- Task categories
- Benchmarking, Agent evaluation
- Formats
- JSONL, Git bundle
- Last updated
- 2026-08-14
Intended uses
- Evaluation & benchmarking
Excluded uses
- Training on evaluation items
- Public redistribution
Provenance & review
- Source summary
Example only: tasks authored by contracted engineers with assignment of rights.
- Rights review
- Not started — Demo listing — no review performed.
- Quality checks
- Example: solution verified by independent engineer; Example: contamination screen
Known limitations
- Demo data — not available for licensing.
Licensing conditions
- Example: evaluation-only licence; no training use.
Final terms are set in a written licence. See buyer licensing terms.
Sample preview
No public preview. A time-limited sample is shared once we approve your request.